Merge 099dcca6c4 into 4d05f84325

* no space after @ in decorator

test/test_jsinterp.py

@@ -577,9 +577,11 @@ class TestJSInterpreter(unittest.TestCase):
     def test_unary_operators(self):
         jsi = JSInterpreter('function f(){return 2  -  - - 2;}')
         self.assertEqual(jsi.call_function('f'), 0)
-        # fails
-        # jsi = JSInterpreter('function f(){return 2 + - + - - 2;}')
-        # self.assertEqual(jsi.call_function('f'), 0)
+        jsi = JSInterpreter('function f(){return 2 + - + - - 2;}')
+        self.assertEqual(jsi.call_function('f'), 0)
+        # https://github.com/ytdl-org/youtube-dl/issues/32815
+        jsi = JSInterpreter('function f(){return 0  - 7 * - 6;}')
+        self.assertEqual(jsi.call_function('f'), 42)
 
     """ # fails so far
     def test_packed(self):

test/test_youtube_signature.py

@@ -158,6 +158,10 @@ _NSIG_TESTS = [
         'https://www.youtube.com/s/player/b7910ca8/player_ias.vflset/en_US/base.js',
         '_hXMCwMt9qE310D', 'LoZMgkkofRMCZQ',
     ),
+    (
+        'https://www.youtube.com/s/player/590f65a6/player_ias.vflset/en_US/base.js',
+        '1tm7-g_A9zsI8_Lay_', 'xI4Vem4Put_rOg',
+    ),
 ]
 
 

youtube_dl/extractor/common.py

@@ -3033,7 +3033,6 @@ class InfoExtractor(object):
             transform_source=transform_source, default=None)
 
     def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
-
         # allow passing `transform_source` through to _find_jwplayer_data()
         transform_source = kwargs.pop('transform_source', None)
         kwfind = compat_kwargs({'transform_source': transform_source}) if transform_source else {}

youtube_dl/extractor/palcomp3.py

@@ -8,7 +8,7 @@ from ..compat import compat_str
 from ..utils import (
     int_or_none,
     str_or_none,
-    try_get,
+    traverse_obj,
 )
 
 
@@ -109,7 +109,7 @@ class PalcoMP3ArtistIE(PalcoMP3BaseIE):
     }
     name'''
 
-    @ classmethod
+    @classmethod
     def suitable(cls, url):
         return False if re.match(PalcoMP3IE._VALID_URL, url) else super(PalcoMP3ArtistIE, cls).suitable(url)
 
@@ -118,7 +118,8 @@ class PalcoMP3ArtistIE(PalcoMP3BaseIE):
         artist = self._call_api(artist_slug, self._ARTIST_FIELDS_TMPL)['artist']
 
         def entries():
-            for music in (try_get(artist, lambda x: x['musics']['nodes'], list) or []):
+            for music in traverse_obj(artist, (
+                    'musics', 'nodes', lambda _, m: m['musicID'])):
                 yield self._parse_music(music)
 
         return self.playlist_result(
@@ -137,7 +138,7 @@ class PalcoMP3VideoIE(PalcoMP3BaseIE):
             'title': 'Maiara e Maraisa - Você Faz Falta Aqui - DVD Ao Vivo Em Campo Grande',
             'description': 'md5:7043342c09a224598e93546e98e49282',
             'upload_date': '20161107',
-            'uploader_id': 'maiaramaraisaoficial',
+            'uploader_id': '@maiaramaraisaoficial',
             'uploader': 'Maiara e Maraisa',
         }
     }]

youtube_dl/extractor/twitter.py

@@ -193,6 +193,22 @@ class TwitterIE(TwitterBaseIE):
     _VALID_URL = TwitterBaseIE._BASE_REGEX + r'(?:(?:i/web|[^/]+)/status|statuses)/(?P<id>\d+)'
 
     _TESTS = [{
+        'url': 'https://twitter.com/TulsiGabbard/status/1555878318469091330',
+        'md5': '8c45b02c4f324e0b9054ceb888091283',
+        'info_dict': {
+            'id': '1555878318469091330',
+            'ext': 'mp4',
+            'title': 'Tulsi Gabbard 🌺 - Puberty-blocking ...',
+            'description': 'Puberty-blocking procedures promoted by the Biden/Harris Admin are child abuse. The FDA has recently confirmed these hormones/drugs have extremely dangerous side effects, like brain swelling and vision loss. https://t.co/pcc70Uyt76',
+            'thumbnail': r're:^https?://.*\.jpg',
+            'uploader': 'Tulsi Gabbard 🌺',
+            'uploader_id': 'TulsiGabbard',
+            'duration': 44.466,
+            'timestamp': 1659785272,
+            'upload_date': '20220806',
+            'age_limit': 0,
+        },
+    }, {
         'url': 'https://twitter.com/freethenipple/status/643211948184596480',
         'info_dict': {
             'id': '643211948184596480',
@@ -213,7 +229,7 @@ class TwitterIE(TwitterBaseIE):
         'info_dict': {
             'id': '657991469417025536',
             'ext': 'mp4',
-            'title': 'Gifs - tu vai cai tu vai cai tu nao eh capaz disso tu vai cai',
+            'title': 'Gifs - "tu vai cai tu vai cai tu nao eh capaz disso[...]',
             'description': 'Gifs on Twitter: "tu vai cai tu vai cai tu nao eh capaz disso tu vai cai https://t.co/tM46VHFlO5"',
             'thumbnail': r're:^https?://.*\.png',
             'uploader': 'Gifs',
@@ -226,10 +242,10 @@ class TwitterIE(TwitterBaseIE):
         'info_dict': {
             'id': '665052190608723968',
             'ext': 'mp4',
-            'title': 'Star Wars - A new beginning is coming December 18. Watch the official 60 second #TV spot for #StarWars: #TheForceAwakens.',
+            'title': 'Star Wars | Andor Premieres Sept 21 on Disney+ - A new beginning is coming December ...',
             'description': 'A new beginning is coming December 18. Watch the official 60 second #TV spot for #StarWars: #TheForceAwakens. https://t.co/OkSqT2fjWJ',
             'uploader_id': 'starwars',
-            'uploader': 'Star Wars',
+            'uploader': 'Star Wars | Andor Premieres Sept 21 on Disney+',
             'timestamp': 1447395772,
             'upload_date': '20151113',
         },
@@ -238,7 +254,7 @@ class TwitterIE(TwitterBaseIE):
         'info_dict': {
             'id': '705235433198714880',
             'ext': 'mp4',
-            'title': "Brent Yarina - Khalil Iverson's missed highlight dunk. And made highlight dunk. In one highlight.",
+            'title': 'Brent Yarina - Khalil Iverson\'s missed highlight dunk. And made highlight dunk[...]',
             'description': "Khalil Iverson's missed highlight dunk. And made highlight dunk. In one highlight. https://t.co/OrxcJ28Bns",
             'uploader_id': 'BTNBrentYarina',
             'uploader': 'Brent Yarina',
@@ -255,11 +271,11 @@ class TwitterIE(TwitterBaseIE):
         'info_dict': {
             'id': '700207533655363584',
             'ext': 'mp4',
-            'title': 'simon vertugo - BEAT PROD: @suhmeduh #Damndaniel',
+            'title': 'jaydin donte geer - BEAT PROD: @suhmeduh #Damndaniel',
             'description': 'BEAT PROD: @suhmeduh  https://t.co/HBrQ4AfpvZ #Damndaniel https://t.co/byBooq2ejZ',
             'thumbnail': r're:^https?://.*\.jpg',
-            'uploader': 'simon vertugo',
-            'uploader_id': 'simonvertugo',
+            'uploader': 'jaydin donte geer',
+            'uploader_id': 'jaydingeer',
             'duration': 30.0,
             'timestamp': 1455777459,
             'upload_date': '20160218',
@@ -282,7 +298,7 @@ class TwitterIE(TwitterBaseIE):
         'info_dict': {
             'id': '719944021058060289',
             'ext': 'mp4',
-            'title': 'Captain America - @King0fNerd Are you sure you made the right choice? Find out in theaters.',
+            'title': 'Captain America - @King0fNerd ...',
             'description': '@King0fNerd Are you sure you made the right choice? Find out in theaters. https://t.co/GpgYi9xMJI',
             'uploader_id': 'CaptainAmerica',
             'uploader': 'Captain America',
@@ -322,7 +338,7 @@ class TwitterIE(TwitterBaseIE):
         'info_dict': {
             'id': '910031516746514432',
             'ext': 'mp4',
-            'title': 'Préfet de Guadeloupe - [Direct] #Maria Le centre se trouve actuellement au sud de Basse-Terre. Restez confinés. Réfugiez-vous dans la pièce la + sûre.',
+            'title': 'Préfet de Guadeloupe - [Direct] #Maria Le centre se trouve...',
             'thumbnail': r're:^https?://.*\.jpg',
             'description': '[Direct] #Maria Le centre se trouve actuellement au sud de Basse-Terre. Restez confinés. Réfugiez-vous dans la pièce la + sûre. https://t.co/mwx01Rs4lo',
             'uploader': 'Préfet de Guadeloupe',
@@ -340,7 +356,7 @@ class TwitterIE(TwitterBaseIE):
         'info_dict': {
             'id': '1001551623938805763',
             'ext': 'mp4',
-            'title': 're:.*?Shep is on a roll today.*?',
+            'title': 'Lis Power - Holy shit, Shep is on a roll today....',
             'thumbnail': r're:^https?://.*\.jpg',
             'description': 'md5:37b9f2ff31720cef23b2bd42ee8a0f09',
             'uploader': 'Lis Power',
@@ -357,7 +373,7 @@ class TwitterIE(TwitterBaseIE):
         'info_dict': {
             'id': '1087791357756956680',
             'ext': 'mp4',
-            'title': 'Twitter - A new is coming.  Some of you got an opt-in to try it now. Check out the emoji button, quick keyboard shortcuts, upgraded trends, advanced search, and more. Let us know your thoughts!',
+            'title': 'Twitter - A new is coming.  Some of you got a...',
             'thumbnail': r're:^https?://.*\.jpg',
             'description': 'md5:6dfd341a3310fb97d80d2bf7145df976',
             'uploader': 'Twitter',
@@ -383,7 +399,7 @@ class TwitterIE(TwitterBaseIE):
         'info_dict': {
             'id': '1349794411333394432',
             'ext': 'mp4',
-            'title': 'md5:d1c4941658e4caaa6cb579260d85dcba',
+            'title': 'Brooklyn Nets - WATCH: Sean Marks\' full media sessi...',
             'thumbnail': r're:^https?://.*\.jpg',
             'description': 'md5:71ead15ec44cee55071547d6447c6a3e',
             'uploader': 'Brooklyn Nets',
@@ -440,14 +456,24 @@ class TwitterIE(TwitterBaseIE):
                 'tweet_mode': 'extended',
             })
 
-        title = description = status['full_text'].replace('\n', ' ')
-        # strip  'https -_t.co_BJYgOjSeGA' junk from filenames
-        title = re.sub(r'\s+(https?://[^ ]+)', '', title)
+        # Get metadata
+        description = status['full_text'].replace('\n', ' ')
         user = status.get('user') or {}
         uploader = user.get('name')
+        uploader_id = user.get('screen_name')
+        timestamp = unified_timestamp(status.get('created_at'))
+
+        # Keep the file name short so it doesn't exceed filesystem limits
+        title = re.sub(r'\s+(https?://[^ ]+)', '', description)
+        if len(title) > 40:
+            trim = 35
+            # allow up to 10 more characters to find whitespace
+            m = re.search(r'\s+', title)
+            if m and m.start() > 10:
+                trim = m.start() + 1
+            title = title[:trim] + '...'
         if uploader:
             title = '%s - %s' % (uploader, title)
-        uploader_id = user.get('screen_name')
 
         tags = []
         for hashtag in (try_get(status, lambda x: x['entities']['hashtags'], list) or []):
@@ -461,7 +487,7 @@ class TwitterIE(TwitterBaseIE):
             'title': title,
             'description': description,
             'uploader': uploader,
-            'timestamp': unified_timestamp(status.get('created_at')),
+            'timestamp': timestamp,
             'uploader_id': uploader_id,
             'uploader_url': 'https://twitter.com/' + uploader_id if uploader_id else None,
             'like_count': int_or_none(status.get('favorite_count')),

youtube_dl/jsinterp.py

@@ -14,6 +14,7 @@ from .utils import (
     remove_quotes,
     unified_timestamp,
     variadic,
+    write_string,
 )
 from .compat import (
     compat_basestring,
@@ -53,15 +54,16 @@ def wraps_op(op):
 
 # NB In principle NaN cannot be checked by membership.
 # Here all NaN values are actually this one, so _NaN is _NaN,
-# although _NaN != _NaN.
+# although _NaN != _NaN. Ditto Infinity.
 
 _NaN = float('nan')
+_Infinity = float('inf')
 
 
 def _js_bit_op(op):
 
     def zeroise(x):
-        return 0 if x in (None, JS_Undefined, _NaN) else x
+        return 0 if x in (None, JS_Undefined, _NaN, _Infinity) else x
 
     @wraps_op(op)
     def wrapped(a, b):
@@ -84,7 +86,7 @@ def _js_arith_op(op):
 def _js_div(a, b):
     if JS_Undefined in (a, b) or not (a or b):
         return _NaN
-    return operator.truediv(a or 0, b) if b else float('inf')
+    return operator.truediv(a or 0, b) if b else _Infinity
 
 
 def _js_mod(a, b):
@@ -220,6 +222,42 @@ class LocalNameSpace(ChainMap):
         return 'LocalNameSpace%s' % (self.maps, )
 
 
+class Debugger(object):
+    ENABLED = False
+
+    @staticmethod
+    def write(*args, **kwargs):
+        level = kwargs.get('level', 100)
+
+        def truncate_string(s, left, right=0):
+            if s is None or len(s) <= left + right:
+                return s
+            return '...'.join((s[:left - 3], s[-right:] if right else ''))
+
+        write_string('[debug] JS: {0}{1}\n'.format(
+            '  ' * (100 - level),
+            ' '.join(truncate_string(compat_str(x), 50, 50) for x in args)))
+
+    @classmethod
+    def wrap_interpreter(cls, f):
+        def interpret_statement(self, stmt, local_vars, allow_recursion, *args, **kwargs):
+            if cls.ENABLED and stmt.strip():
+                cls.write(stmt, level=allow_recursion)
+            try:
+                ret, should_ret = f(self, stmt, local_vars, allow_recursion, *args, **kwargs)
+            except Exception as e:
+                if cls.ENABLED:
+                    if isinstance(e, ExtractorError):
+                        e = e.orig_msg
+                    cls.write('=> Raises:', e, '<-|', stmt, level=allow_recursion)
+                raise
+            if cls.ENABLED and stmt.strip():
+                if should_ret or not repr(ret) == stmt:
+                    cls.write(['->', '=>'][should_ret], repr(ret), '<-|', stmt, level=allow_recursion)
+            return ret, should_ret
+        return interpret_statement
+
+
 class JSInterpreter(object):
     __named_object_counter = 0
 
@@ -307,8 +345,7 @@ class JSInterpreter(object):
     def __op_chars(cls):
         op_chars = set(';,[')
         for op in cls._all_operators():
-            for c in op[0]:
-                op_chars.add(c)
+            op_chars.update(op[0])
         return op_chars
 
     def _named_object(self, namespace, obj):
@@ -326,9 +363,8 @@ class JSInterpreter(object):
         # collections.Counter() is ~10% slower in both 2.7 and 3.9
         counters = dict((k, 0) for k in _MATCHING_PARENS.values())
         start, splits, pos, delim_len = 0, 0, 0, len(delim) - 1
-        in_quote, escaping, skipping = None, False, 0
-        after_op, in_regex_char_group = True, False
-
+        in_quote, escaping, after_op, in_regex_char_group = None, False, True, False
+        skipping = 0
         for idx, char in enumerate(expr):
             paren_delta = 0
             if not in_quote:
@@ -382,10 +418,12 @@ class JSInterpreter(object):
         return separated[0][1:].strip(), separated[1].strip()
 
     @staticmethod
-    def _all_operators():
-        return itertools.chain(
-            # Ref: https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Operators/Operator_Precedence
-            _SC_OPERATORS, _LOG_OPERATORS, _COMP_OPERATORS, _OPERATORS)
+    def _all_operators(_cached=[]):
+        if not _cached:
+            _cached.extend(itertools.chain(
+                # Ref: https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Operators/Operator_Precedence
+                _SC_OPERATORS, _LOG_OPERATORS, _COMP_OPERATORS, _OPERATORS))
+        return _cached
 
     def _operator(self, op, left_val, right_expr, expr, local_vars, allow_recursion):
         if op in ('||', '&&'):
@@ -416,7 +454,7 @@ class JSInterpreter(object):
         except Exception as e:
             if allow_undefined:
                 return JS_Undefined
-            raise self.Exception('Cannot get index {idx:.100}'.format(**locals()), expr=repr(obj), cause=e)
+            raise self.Exception('Cannot get index {idx!r:.100}'.format(**locals()), expr=repr(obj), cause=e)
 
     def _dump(self, obj, namespace):
         try:
@@ -438,6 +476,7 @@ class JSInterpreter(object):
     _FINALLY_RE = re.compile(r'finally\s*\{')
     _SWITCH_RE = re.compile(r'switch\s*\(')
 
+    @Debugger.wrap_interpreter
     def interpret_statement(self, stmt, local_vars, allow_recursion=100):
         if allow_recursion < 0:
             raise self.Exception('Recursion limit reached')
@@ -511,7 +550,6 @@ class JSInterpreter(object):
                 expr = self._dump(inner, local_vars) + outer
 
         if expr.startswith('('):
-
             m = re.match(r'\((?P<d>[a-z])%(?P<e>[a-z])\.length\+(?P=e)\.length\)%(?P=e)\.length', expr)
             if m:
                 # short-cut eval of frequently used `(d%e.length+e.length)%e.length`, worth ~6% on `pytest -k test_nsig`
@@ -693,7 +731,7 @@ class JSInterpreter(object):
                 (?P<op>{_OPERATOR_RE})?
                 =(?!=)(?P<expr>.*)$
             )|(?P<return>
-                (?!if|return|true|false|null|undefined)(?P<name>{_NAME_RE})$
+                (?!if|return|true|false|null|undefined|NaN|Infinity)(?P<name>{_NAME_RE})$
             )|(?P<indexing>
                 (?P<in>{_NAME_RE})\[(?P<idx>.+)\]$
             )|(?P<attribute>
@@ -727,11 +765,12 @@ class JSInterpreter(object):
             raise JS_Break()
         elif expr == 'continue':
             raise JS_Continue()
-
         elif expr == 'undefined':
             return JS_Undefined, should_return
         elif expr == 'NaN':
             return _NaN, should_return
+        elif expr == 'Infinity':
+            return _Infinity, should_return
 
         elif md.get('return'):
             return local_vars[m.group('name')], should_return
@@ -760,18 +799,28 @@ class JSInterpreter(object):
             right_expr = separated.pop()
             # handle operators that are both unary and binary, minimal BODMAS
             if op in ('+', '-'):
+                # simplify/adjust consecutive instances of these operators
                 undone = 0
                 while len(separated) > 1 and not separated[-1].strip():
                     undone += 1
                     separated.pop()
                 if op == '-' and undone % 2 != 0:
                     right_expr = op + right_expr
+                elif op == '+':
+                    while len(separated) > 1 and separated[-1].strip() in self.OP_CHARS:
+                        right_expr = separated.pop() + right_expr
+                # hanging op at end of left => unary + (strip) or - (push right)
                 left_val = separated[-1]
                 for dm_op in ('*', '%', '/', '**'):
                     bodmas = tuple(self._separate(left_val, dm_op, skip_delims=skip_delim))
                     if len(bodmas) > 1 and not bodmas[-1].strip():
                         expr = op.join(separated) + op + right_expr
-                        right_expr = None
+                        if len(separated) > 1:
+                            separated.pop()
+                            right_expr = op.join((left_val, right_expr))
+                        else:
+                            separated = [op.join((left_val, right_expr))]
+                            right_expr = None
                         break
                 if right_expr is None:
                     continue
@@ -797,6 +846,8 @@ class JSInterpreter(object):
 
             def eval_method():
                 if (variable, member) == ('console', 'debug'):
+                    if Debugger.ENABLED:
+                        Debugger.write(self.interpret_expression('[{}]'.format(arg_str), local_vars, allow_recursion))
                     return
                 types = {
                     'String': compat_str,

youtube_dl/utils.py

@@ -2406,7 +2406,7 @@ class ExtractorError(YoutubeDLError):
         """ tb, if given, is the original traceback (so that it can be printed out).
         If expected is set, this is a normal error message and most likely not a bug in youtube-dl.
         """
-
+        self.orig_msg = msg
         if sys.exc_info()[0] in (compat_urllib_error.URLError, socket.timeout, UnavailableVideoError):
             expected = True
         if video_id is not None: